Identification Of Wear Particles Research Articles

An oil wear particle identification method based on Wasserstein generative adversarial network and improved CNN using a custom-built optical imaging sensor

The accurate classification of wear particles in lubricating oil holds significant importance for understanding the wear characteristics of lubricated components. However, there are prominent issues, such as blurred particle images, limited identification efficiency, and low generalization capabilities. To solve such issues, this research initially devises a novel online monitoring sensor to collect particle images based on the optical method. Subsequently, an improved Wasserstein Generative Adversarial Network (WGAN) is developed to ameliorate the categorization performance, which is hampered by data imbalance. Lastly, a MobileNetV2-SENet-based Convolutional Neural Network (MSCNN) is proposed to autonomously discover the associations between images and types of wear particles. Experimental results illustrate that our method achieves a recognition accuracy rate of 92.5%. The analysis outcomes between several comparative networks show that our approach can effectively improve recognition performance when limited samples are available. This work offers an effective solution for automatically identifying wear particles in oil and can be a powerful predictive maintenance tool for rotating mechanical equipment.

Research on an Intelligent Classification Algorithm of Ferrography Wear Particles Based on Integrated ResNet50 and SepViT

The wear particle classification algorithm proposed is based on an integrated ResNet50 and Vision Transformer, aiming to address the problems of a complex background, overlapping and similar characteristics of wear particles, low classification accuracy, and the difficult identification of small target wear particles in the region. Firstly, an ESRGAN algorithm is used to improve image resolution, and then the Separable Vision Transformer (SepViT) is introduced to replace ViT. The ResNet50-SepViT model (SV-ERnet) is integrated by combining the ResNet50 network with SepViT through weighted soft voting, enabling the intelligent identification of wear particles through transfer learning. Finally, in order to reveal the action mechanism of SepViT, the different abrasive characteristics extracted by the SepViT model are visually explained using the Grad-CAM visualization method. The experimental results show that the proposed integrated SV-ERnet has a high recognition rate and robustness, with an accuracy of 94.1% on the test set. This accuracy is 1.8%, 6.5%, 4.7%, 4.4%, and 6.8% higher than that of ResNet101, VGG16, MobileNetV2, AlexNet, and EfficientV1, respectively; furthermore, it was found that the optimal weighting factors are 0.5 and 0.5.

An automated system for polymer wear debris analysis in total disc arthroplasty using convolution neural network.

Introduction: Polymer wear debris is one of the major concerns in total joint replacements due to wear-induced biological reactions which can lead to osteolysis and joint failure. The wear-induced biological reactions depend on the wear volume, shape and size of the wear debris and their volumetric concentration. The study of wear particles is crucial in analysing the failure modes of the total joint replacements to ensure improved designs and materials are introduced for the next generation of devices. Existing methods of wear debris analysis follow a traditional approach of computer-aided manual identification and segmentation of wear debris which encounters problems such as significant manual effort, time consumption, low accuracy due to user errors and biases, and overall lack of insight into the wear regime. Methods: This study proposes an automatic particle segmentation algorithm using adaptive thresholding followed by classification using Convolution Neural Network (CNN) to classify ultra-high molecular weight polyethylene polymer wear debris generated from total disc replacements tested in a spine simulator. A CNN takes object pixels as numeric input and uses convolution operations to create feature maps which are used to classify objects. Results: Classification accuracies of up to 96.49% were achieved for the identification of wear particles. Particle characteristics such as shape, size and area were estimated to generate size and volumetric distribution graphs. Discussion: The use of computer algorithms and CNN facilitates the analysis of a wider range of wear debris with complex characteristics with significantly fewer resources which results in robust size and volume distribution graphs for the estimation of the osteolytic potential of devices using functional biological activity estimates.

The result greyness problem of the grey relational analysis and its solution

This paper discusses how to deal with the greyness problem in the system from the perspective of “result”. Aiming at the greyness problem of the traditional grey relational analysis result, an information fusion grey relational analysis method based on D-S evidence theory and multi solution information fusion is proposed, which mends the traditional grey relational analysis method. The results show that the method proposed in the study has better effect than the traditional grey relational analysis method, and has higher accuracy in the wear particle identification, which indicates that it can further expand the application scope of the grey relational analysis method.

A New Transfer Learning Ensemble Model with New Training Methods for Gear Wear Particle Recognition

Aiming at solving the acquisition problems of wear particle data of large-modulus gear teeth and few training datasets, an integrated model of LCNNE based on transfer learning is proposed in this paper. Firstly, the wear particles are diagnosed and classified by connecting a new joint loss function and two pretrained models VGG19 and GoogLeNet. Subsequently, the wear particles in gearbox lubricating oil are chosen as the experimental object to make a comparison. Compared with the other four models’ experimental results, the model superiority in wear particle identification and classification is verified. Taking five models as feature extractors and support vector machines as classifiers, the experimental results and comparative analysis reveal that the LCNNE model is better than the other four models because its feature expression ability is stronger than that of the other four models.

Texture extraction of wear particles based on improved random Hough transform and visual saliency

Texture recognition is very important and necessary for the identification of wear particles. In order to extract distinctive textures from the images of wear particles, a new texture extraction method based on improved random Hough transform and visual saliency of wear particles is proposed in this paper. This method mimics what human eyes see when they analyze wear particle images. Firstly, Canny detector is used to obtain the initial texture seed image. Secondly, the improved random Hough transform based on the connectivity of pixels is applied on texture seed images to extract the texture primitives, such as lines or circles, which represent visual saliency of the texture on wear particles. Finally, through the statistics of texture primitives, the regularity and randomness of texture pattern could be analyzed so as to determine the type of wear particles. Experimental results show that this method is effective to distinguish the severe sliding, fatigue wear particles and black oxides, it is an intuitive and fast method.

Three-dimensional reconstruction of wear particle surface based on photometric stereo

Ferrography has emerged as a significant candidate for the provision of relevant information for the determination of the status of machine wear. However, this conventional methodology provides only marginal results due to its inability to provide 3-dimensinal (3D) surface information. The authors of this paper have developed a methodology based on photometric-stereo towards the enhancement of the capabilities of ferrography. This enhanced and innovative methodology consists of three main components, the multi-illumination image acquisition, the wear particle extraction, and the 3D surface reconstruction. The methodology ensures the reliable and efficient extraction of the wear particles surface topographies for further feature-based wear particle identification. The performance of this methodology has been compared with results observed from the laser scanning confocal microscopy. The outcome of this comparison has depicted that the methodology involving this new low-cost ferrography system exhibits very high accuracy for the 3D surface feature extraction of wear particles.

Diagnostic guidelines for the histological particle algorithm in the periprosthetic neo-synovial tissue

BackgroundThe identification of implant wear particles and non-implant related particles and the characterization of the inflammatory responses in the periprosthetic neo-synovial membrane, bone, and the synovial-like interface membrane (SLIM) play an important role for the evaluation of clinical outcome, correlation with radiological and implant retrieval studies, and understanding of the biological pathways contributing to implant failures in joint arthroplasty. The purpose of this study is to present a comprehensive histological particle algorithm (HPA) as a practical guide to particle identification at routine light microscopy examination.MethodsThe cases used for particle analysis were selected retrospectively from the archives of two institutions and were representative of the implant wear and non-implant related particle spectrum. All particle categories were described according to their size, shape, colour and properties observed at light microscopy, under polarized light, and after histochemical stains when necessary. A unified range of particle size, defined as a measure of length only, is proposed for the wear particles with five classes for polyethylene (PE) particles and four classes for conventional and corrosion metallic particles and ceramic particles.ResultsAll implant wear and non-implant related particles were described and illustrated in detail by category. A particle scoring system for the periprosthetic tissue/SLIM is proposed as follows: 1) Wear particle identification at light microscopy with a two-step analysis at low (× 25, × 40, and × 100) and high magnification (× 200 and × 400); 2) Identification of the predominant wear particle type with size determination; 3) The presence of non-implant related endogenous and/or foreign particles. A guide for a comprehensive pathology report is also provided with sections for macroscopic and microscopic description, and diagnosis.ConclusionsThe HPA should be considered a standard for the histological analysis of periprosthetic neo-synovial membrane, bone, and SLIM. It provides a basic, standardized tool for the identification of implant wear and non-implant related particles at routine light microscopy examination and aims at reducing intra-observer and inter-observer variability to provide a common platform for multicentric implant retrieval/radiological/histological studies and valuable data for the risk assessment of implant performance for regional and national implant registries and government agencies.

A hybrid search-tree discriminant technique for multivariate wear debris classification

Wear debris analysis (WDA) is an effective machine health monitoring technique in which wear debris attributes can reveal wear mechanisms and particle colours can be used to detect oxidation. However, most of the existing approaches to WDA are based on analytical ferrography. Online wear debris typing is still difficult to achieve because of the low resolution of online particle images. In this work, a hybrid search-tree discriminant technique is described. It permits identification of online wear particles by combining their colour attributes and multi-view particle features. A multi-class support vector data description multi-SVDD, the K-means and a support vector machine SVM are integrated to establish a three-level search-tree model to discriminate multivariate wear debris including red and black oxides, cutting, spherical, fatigue and sliding particles. First, the oxides are identified based on their special colour information using the multi-SVDD. Second, the K-means is used to look for the clustering centres of cutting and spherical particles by utilizing their distinct features of aspect ratio and sphericity. The third level built by SVM is adopted to distinguish fatigue and sliding particles based on their height aspect ratio and height-to-width aspect ratio. Experiments have demonstrated that the search-tree discrimination model is effective for multivariate wear debris classification. The proposed method provides a solution to the existing problems in online wear particle identification for wear mechanism analysis.

Histopathological diagnostic work-up of joint endoprosthesis-associated pathologies

Increasing classes of joint implants and the combination of materials results in increased and wear-associated pathologies. According to the revised consensus classification, the following types can be recognized at conventional histological examination: TypeI, particle-induced type; TypeII, infection type; TypeIII, combination type; TypeIV, indifferent type; TypeV arthrofibrotic type; TypeVI, allergic/immunological/toxic adverse reactions and TypeVII, bone pathologies. Wear particles are histopathologically characterized according to the Krenn particle algorithm which focuses on adescriptive identification of wear particles and the differentiation of other nonwear-related particles. TypeVII is considered histologically when there is evidence of aperivascular/interstitial lymphocytic CD20- and CD3-positive infiltrate, presence of mast cells and eosinophils, and tissue necrosis/infarction associated with implant wear material. Since wear particle-induced toxicity cannot be differentiated with certainty from hypersensitivity/allergic reaction on histological examination, immunological-allergological and clinical data should be used as supplementary criteria for the differential diagnosis. Tissue sampling should be performed from periprosthetic soft tissue with location mapping and when feasible also from bone tissue. Additional information regarding the type of implant and clinical, radiological, immunological, and microbiology data should be available to the pathologist. Further immunohistochemical studies are recommended in the following settings: infection (CD15, CD20, CD68); prosthesis-associated arthrofibrosis (β‑catenin); allergic/immunologic/toxic adverse reactions (CD20, CD3, CD4, CD8, CD117 and for T‑cell characterization T‑bet, GATA-3, and FOXP3).

An experiment on wear particle’s texture analysis and identification by using deterministic tourist walk algorithm

Purpose – This study aims to use a deterministic tourist walk to build a system that can identify wear particles. Wear particles provide detailed information about the wear processes taking place between mechanical components. Identification of the type of wear particles by image processing and pattern recognition is key to effective online monitoring algorithm. There are three kinds of particles that are particularly difficult to distinguish: severe sliding wear particles, fatigue spall particles and laminar particles. Design/methodology/approach – In this study, an identification method is tested using the deterministic tourist walking (DTW) method. This study examined whether this algorithm can be used in particle identification. If it does, can it outperform the traditional texture analysis methods such as Discrete wavelet transform or co-occurrence matrix. Different parameters such as walk’s memory size, size of image samples, different inputting vectors and different classifiers were compared. Findings – The DTW algorithm showed promising result compared to traditional texture extraction methods: discrete wavelet transform and co-occurrence matrix. The DTW method offers a higher identification accuracy and a simple feature vector. A conclusion can be drawn that the DTW method is suited for particle identification and can be put into practical use in condition monitoring systems. Originality/value – This paper combined DTW algorithm with wear particle identification problem.

A wear particle identification method by combining principal component analysis and grey relational analysis

The process to identify wear particles concerns a variety of parameters, some of which may be redundant, and influences the efficiency of computer image analysis. In order to improve the accuracy and speed of debris identification, this paper proposes a new algorithm that combines principal component analysis and grey relational analysis (CPGA). First, principal component analysis is used to optimise the characteristic parameters of wear particles. Then, an improved grey relational analysis is used to discriminate between similar types of wear particles, such as severe sliding and fatigue particles. The experimental results indicate that the CPGA algorithm can successfully solve the information redundancy problem resulting from multiple parameters and proves to be a practical method to identify wear particles quickly and accurately.

Classification of Wear Debris Using Weighted Fuzzy Cluster Method

The analysis and identification of wear particles for machine condition monitoring is usually very time-consuming by experienced inspectors. In order to remedy the limitation, automation of the analysis procedure appears to be necessary. A novel weighted fuzzy c-means algorithm for wear particle classification is proposed in this paper. The algorithm uses the variation of the pixel intensities of a region to choose strong resembling area. Then, the spatial relationships of the membership function are constructed to regulate the pixel membership obtained from the FCM object function. Finally, wear debris are classified based on the fuzzy membership. The example shows that the method is briefly and effectively.

Recognition of texture types of wear particles

Microscopic wear particles are produced in all machines containing moving parts in contact. The particles, transported by a lubricant from wear sites; carry important information relating to the condition of the machinery. This information is classified by compositional and six morphological attributes of particle size, shape, edge details, color, thickness ratio, and surface texture. This article describes an automated system for surface texture identification of wear particles by using artificial neural networks. The aim is to classify these particles according to their morphological attributes and by using the information obtained, to predict wear failure modes in engines and other machinery. This approach will enable the manufacturing industry to improve quality, productivity and economy. The procedure reported in this article is based on gray level co-occurrence matrices which are used to train a feed-forward neural network classifier in order to distinguish among seven different patterns which can aid in the identification of wear particles. The investigated patterns are: smooth, rough, striations, holes, pitted, cracked, and serrated. An accuracy classification rate of 94.64% has been achieved and is shown by a confusion matrix.

A Web-Based Remote Intelligent Expert System for Ferrography Diagnosis

Abstract. A ferrography expert system for wear particle analysi s can be an effective means of condition monitoring and fault diagnosis. The knowledge of the ferrography diagnosis expert system cannot be represented by certain rules because it is f uzzy and random, therefore an intelligent expert system is developed. A neural network with self -deleting nodes is employed in the intelligent expert system, and this overcomes to some degree the disadvantages of a traditional expert system and makes it possible to automatically acquire knowl edge by learning from samples and to realize entirely automatic processing, from wear particle rec ognition to wear diagnosis. Introduction Wear is one of the main factors causing breakdown and faults in mac hines. Therefore the ferrography technique for wear particle analysis can be an effective way for condition monitoring and fault diagnosis [1,2]. It is difficult for the ferrography tec hnique to be widely used because it mainly depends on operator’s experience. Therefore it is very useful to develop a ferrography diagnosis expert system. Some expert systems have been develope d in this field and led to some effective results. FAST, FAST PLUS [3,4], CASPA [5] and CAVE [6,7] are examples which employ traditional expert systems. It is difficult to define the knowledge of the ferrography diagnosis expert system by certain rules, because it is fuzz y and random. Some intelligent expert systems were studied [8,9], but they were still not fully automatic integrated systems. A web-based remote intelligent expert system for ferrogr aphy diagnosis (RIESFD), based on a neural network, fuzzy theory and Internet for wear particle ident ification. Wear diagnosis is presented in this paper. RIESFD overcomes to some degree the disadvanta ges of a traditional expert system, and it can automatically acquire knowledge by learning fr om samples and realize entirely automatic processing from wear particle identification to we ar diagnosis. It is very important for this expert system based on a neural network to be provided with enough samples for its learning, therefore it should be realized to collect all kinds of typical wear particle samples and share the knowledge between expert systems in different locations. The Internet is a good tool for the exchange and contact between different systems. Therefore the appl ication of a remote diagnosis course on Internet is the direction of the ferrography technology development. Basic Principle of RIESFD RIESFD is developed for full auto-identification of wear partic les and wear diagnosis. The system includes three steps: (1) Identify all wear particles on a ferrograph slice one by one, and acquire the information on the classification of wear particles at one time point. (2) Read all history information on wear particles from a database , and form the distribution of different kinds of wear particles with time. (3) Diagnose the wear fault on the basis of the above, and give the result. With the above idea, the system progress is shown as Fig. 1. The key techniques are wear particle identification and wear fault diagnosis, therefore a ne ural network is employed for this purpose. The wear particles are randomly produced during equipment running, and different wear particles can be produced at different states and different friction pai rs, therefore it is very important

Identification of wear particles of joint prostheses in tissues using laser microprobe mass analysis (LAMMA)

The definite identification of wear particles from joint prostheses is of great importance for the development of joint replacement, as the type and quantity of different wear particles gives information on the wear resistance of implant materials. From the types of prostheses nowadays in use polyethylene wear of the sockets, bone cement wear, metallic and ceramic wear can be generated. Whereas polyethylene wear can be easily identified by its bright luminescence in polarized light and its characteristic configuration, the distinction of the small granular wear particles of the bone cement, metal and ceramic by light microscope is difficult. The laser microprobe mass analysis (LAMMA) is a method, which allows the analysis of single light microscopically detectable wear particles in tissues. Not only contrast medium particles of the bone cements (zirconium oxide or barium sulfate) but also metallic and aluminum oxide particles could be definitely identified within the pseudocapsules as well as in regional lymph nodes by LAMMA-analysis, whereby the bone cement wear predominated. In addition, the distinction between organic substances (as blood degradation products), which may appear similar to wear particles in configuration and colour, and the foreign material is also possible with this method.

Wear particle classification in a fuzzy grey system

The analysis and identification of wear particles for machine condition monitoring is usually conducted by experienced inspectors, and, thus, the process is usually very time-consuming. To overcome this obstacle, grey system theory has been applied in this study. The theory of grey systems is a new technique to perform prediction, relational analysis and decision making in many areas. In this paper, the theory of grey relational grades has been used to classify six types of metallic wear debris whose three-dimensional images are acquired from laser scanning confocal microscopy. Their boundary morphology and surface topology are then described by certain numerical parameters. Since the parameters have different levels of significance for different types of wear debris for particle identification, weighting factors of the parameters have been taken into consideration. To determine the weighting factors for the study, fuzzy logic has been applied. This study has demonstrated that a grey system combined with fuzzy logic can be used to classify wear particles satisfactorily.

Computer image analysis for identification of wear particles

The design of a specialized image analyser to investigate wear particles is described. An automatic grey-level threshold selection method identifies particles from the background. The entire image is enhanced through iterative relaxation labelling. A size distribution of particles within the field of view is produced. The shape of isolated particles can be processed to reveal useful quantitative data that is not produced from human visual examination. Border curvature of these particles is examined as a means of automatic particle recognition. Various aspect ratio measurements, including Fourier descriptors, are also considered. These morphological measurements are used to classify particles into established groups. A number of descriptors such as these are used to form the foundations of a diagnostic expert system.

Identification of wear particles in tissue after implantation of different plastic materials

A method is shown by which to identify wear particles of implanted plastic materials in tissue. A Mettler FP2 hot stage was used to determine the melting point of different elastomer particles in normal histological sections between the crossed nicols. By using a microcolorimeter Type MPE (Leitz, Wetzlar, Germany), additional information about the molecular degradation of wear particles was obtained.